Nuclear Fuel Cycle Studies

The U.S. Department of Energy (DOE), along with many other organizations, has been conducting research on the nuclear fuel cycle for over 40 years. In 2001, the DOE began studying fuel cycle systems under the aegis of the Generation IV initiative. In 2003, the system studies program was expanded as part of the Advanced Fuel Cycle Initiative (AFCI). Work continues under the AFCI, now called the Fuel Cycle Research and Development (FCRD) Program (after a brief period under an umbrella initiative called the Global Nuclear Energy Partnership). This work includes annual reports and other documents produced by researchers at Idaho National Lab and Argonne National Lab, and by FCRD-supported university research. A small selection of U.S. fuel cycle reports, many (but not all) produced with full or partial support of the AFCI/FCRD, is described in section 2.2.2 below.

31 2.2.1 Why Study the Nuclear Fuel Cycle: AFCI/FCRD Goals

Policymakers and technologists have produced several iterations of the goal set to which advanced nuclear fuel cycle technologies should aspire. Each time these goals/values are

restated, the theme is the same: the nuclear fuel cycle should be sustainable, imposing as small an impact as possible on resources and nuclear proliferation. To varying extents, every fuel cycle study invokes goals along these lines, and qualitatively or quantitatively discusses the impacts of advanced technologies on the goals.

The current DOE program, FCRD, states explicitly that its objective is to conduct research and help develop sustainable fuel cycles. It states goals for fuel cycle technologies through a definition: “Sustainable fuel cycle options are those that improve uranium resource utilization, maximize energy generation, minimize waste generation, improve safety, and limit proliferation risk.”(U.S. Department of Energy, 2011) Though the program managers did not explicitly name low cost as a goal, they do indicate later in the document that the fuel cycle must be “acceptable to the American public.” As such, the final option must adhere to a cost goal that was defined in the AFCI mission, to “ensure that advanced fuel cycle technologies cause no significant decrease in the economic competitiveness of nuclear electricity.”(Dixon et al., 2008) Indeed, given the current competitive environment in which all nuclear and non-nuclear energy technologies are judged, a fuel cycle that fails to meet this objective is sure never to be

implemented.

2.2.2 Nuclear Fuel Cycle Reports

Among the factors that make fuel cycle decision making difficult is the sheer volume of information available about technological options. Reports on the subject range from the purely qualitative to those based entirely on quantitative models. On the qualitative end, industry has suggested paths forward for fuel cycle evolution by combining statements from experts.(Nuclear energy for the future: Required research and development capabilities - an industry

perspective2008) Similarly, recommendations to government by advisory committees also tend to come in the form of the gathered wisdom of fuel cycle experts, who themselves are familiar with a wide range of studies (e.g. (Martin, Ahearne, & and the Nuclear Energy Advisory Committee, 2008)). Hybrid qualitative/quantitative studies abound as well. One example is the MIT Future of the Nuclear Fuel Cycle report, which includes fuel cycle system modeling for scenario analysis and a separate model for uranium consumption, alongside qualitative sections

32

assessing the pros and cons of various fuel cycle options.(Kazimi, Moniz, Forsberg, & et al., 2011)

Highly quantitative modeling and simulation fuel cycle studies tend to fall into three broad categories: individual facility analyses (such as in-depth modeling of an advanced reactor), static or equilibrium system analyses (usually discussions of system-wide impacts of the fuel cycle, including linkages between reprocessing plants and advanced reactors), and dynamic analyses (fleet-level system analyses that examine the development of the fuel cycle system over time). Each type of study adds crucial information to the field of fuel cycle analysis.

Among the most important single-facility analyses are those that examine advanced reactors. Reactor studies often include calculations performed using core neutronic and thermal-hydraulic models, in order to simulate the performance of reactor fuel and the potential response of the reactor system to design-basis accidents. Examples include numerous MIT theses (see e.g.

(Pope, Driscoll, & Hejzlar, 2004)). One of the most cited and influential studies in the area of fuel cycle analysis is Edward Hoffman’s design of a sodium-cooled fast reactor (SFR).(Hoffman, Yang, & Hill, 2008) Hoffman performed an analysis of core operation, and in the process

defined the isotopic compositions needed to sustain reactor operation for a range of fast reactor conversion ratios. His isotopic vector “recipes” are used in this thesis as well as Idaho National Lab’s fuel cycle simulation called VISION.

The in-depth studies of individual facilities are vital for any advancement of nuclear technology, but they do not provide all of the information needed to make decisions about fuel cycle evolution. Advanced reactors in fact exist within a complex system of reprocessing plants, light water reactors, fabrication plants, and disposal systems, and these all need to be

technologically and economically harmonized in order to close or significantly alter the fuel cycle.

Many system studies focus on collections of multiple facility types. Wigeland and Bauer, for example, consider the repository impact of thermal recycling by looking at the interaction of light-water reactors (LWRs) and aqueous reprocessing facilities.(Wigeland & Bauer, 2004) They are able to quantify the benefit of limited thermal recycle on repository loading in terms of a mass factor increase that can be loaded into Yucca Mountain. They do not make any analysis of fleet characteristics or nuclear power growth, but instead concentrate on a representative LWR

33 and reprocessing facility. In this sense, the analysis can be said to be “static” – it does not

address growth or change over time.

Studies that estimate the costs of advanced recycling also often take static or equilibrium viewpoints. Parsons and De Roo perform a calculation of advanced fuel cycle costs assuming that LWRs and fast reactors are in equilibrium with one another; this means they assume that the numbers of LWRs and fast reactors are attuned to one another such that there is no (or very little) excess fuel passing between them, and also that the isotopic composition of multi-recycled fuel changes very little with each pass through a fast reactor.(De Roo & Parsons, 2009) This

approach enables them to calculate a levelized cost of electricity for the advanced fuel cycle system that can be compared to the levelized cost of nuclear power today.

The third category consists of dynamic fuel cycle studies. Many of these studies are relatively recent, because complex fuel cycle simulations have become gradually more accessible as the speed of computing has increased. The use of dynamic fuel cycle analysis ranges from single-growth rate, single-technology scenarios intended to bolster qualitative arguments, to extensive use of complex codes that include sensitivity analyses.

The codes which are the workhorses of these analyses similarly range from the very simple to the very complex. Among the most complex and highly-developed fuel cycle codes are VISION (Idaho National Laboratory), DANESS (Argonne National Laboratory), and COSI (CEA – the French Atomic Energy Commission). Of the three, COSI may be the most

sophisticated, because it includes linked neutronics models to calculate fuel compositions as it runs. VISION and DANESS are both written on system dynamics platforms (PowerSim and iThink, respectively), and include millions of input options and extensive output data.

Researchers at Idaho National Lab used VISION in 2008 to perform one of the most extensive fuel cycle analyses to date, called the Dynamic Systems Analysis Report for Nuclear Fuel Recycle (DSARR).(Dixon et al., 2008) DSARR’s primary objective is a comparison of three fuel cycle systems: (1) once-through, (2) “1-tier” (LWR spent fuel is sent to immediate fast reactor recycle of all transuranics), and (3) “2-tier” (LWRs recycle spent fuel in a limited

manner, followed by a later introduction of fast reactors). System scenarios are run assuming a 1.75% growth rate in nuclear energy demand, and assuming single introduction dates for each advanced technology. The technology options are then compared in terms of their uranium usage and waste impacts (which is divided into dimensions of dose, heat, and radiotoxicity). The costs

34

of each system are separately estimated using static assumptions, to provide a further metric of comparison. Proliferation concerns are not addressed in the report.

The notions of uncertainty and flexibility are addressed qualitatively in DSARR. The authors suggest three ways in which fuel cycle systems could be flexible, including (1) burning Pu-MOX in LWRs if fuel separations have begun but fast reactors are not ready for deployment, (2) fast reactors could use excess weapons material or enriched uranium if separations are not operational in time, and (3) demonstration plants could be converted upward to commercial-scale plants. Each of these possibilities could prove to be useful pathways for fuel cycle system

evolution, but they all are more like solutions to potential problems than they are options that can be exercised for a range of uncertain futures. Furthermore, they do not address uncertainties in nuclear power growth, costs of advanced technologies, or the possibility that other events could dramatically alter the desired suite of technologies at any given time.

The authors do, however, quantitatively assess the possibility that advanced technologies might be desired (or only available) later in the future. A sensitivity analysis on the date of fast reactor introduction shows that late deployment of FRs actually slightly increases the number of fast reactors that are ultimately built, because the build rate of FRs is so sensitive to the

availability of used LWR fuel. The effect, however, is minimal. These results track with some of the scenarios studied in this thesis, but they are not discussed further in DSARR.

DSARR is extremely successful in accomplishing its objective of describing some canonical technology options for fuel cycle evolution, and comparing them to the stated goals of the AFCI. The authors have made an enormous contribution to the information required by policymakers to move development of the fuel cycle forward. The picture painted by DSARR and many other dynamic fuel cycle studies, however, is incomplete. In order to make good decisions about the future of nuclear power technologies, policymakers will need to better understand the impacts of various uncertainties and the tradeoffs between fuel cycle impacts.

Nearly all studies like DSARR address a range of metrics (e.g. resource sustainability in the form of uranium consumption, heat output of the waste), but do not offer a quantitative assessment of how the objectives of sustainability, cost, safety, and security trade off among one another.

The framework outlined and demonstrated in this thesis forces explicit consideration of sources of uncertainty, the level of flexibility with which various systems address those

uncertainties, and the tradeoffs between system values. The technology options evaluated here

35 are few, because the intent was to demonstrate the method more than to provide a comprehensive fuel cycle option analysis. The hope is that as fuel cycle information becomes richer, new data and important options can be added into the framework in order to provide policymakers with a holistic analysis of fuel cycle evolution pathways. Ultimately, this framework can help identify the most (or least) promising ways to balance competing goals and move fuel cycle development forward.